analysis of effect of impeller geometry including blade

107

ISSN 13921207. MECHANIKA. 2017 Volume 23(1): 107119

Analysis of effect of impeller geometry including blade outlet angle on

the performance of multi-pressure pumps: Simulation and experiment

N. Mohammadi*, M. Fakharzadeh** *Department of Mechanical Engineering, Islamic Azad University, Parand Branch, Parand, Iran,

E-mail: [email protected]

**Department of Mechanical Engineering, Islamic Azad University, Parand Branch, Parand, Iran,

E-mail: [email protected]

http://dx.doi.org/10.5755/j01.mech.23.1.17676

1. Introduction

Today, using pumps, to increase the liquid flow

power, plays an important role in economic arenas as well

as energy generation and water working projects. Centrifu-

gal pumps have found increased applications in different

industries where they have served as heart of a system to

displace some fluid(s); these pumps, however, consume a

large portion of total energy consumed by such systems.

Centrifugal pumps have been widely applied across differ-

ent industries including petrochemical complexes, refiner-

ies, thermal power plants, military industries, nuclear

plants, agriculture and visually every industry where there

may be a need for increased fluid flow pressure. As such,

many researchers have tried to improve the performance of

such pumps.

One of the principle parameters to consider when

choosing a pump is the generable head by the pump which

is a function of different parameters such as impeller diam-

eter, pump rotational speed, pump flow rate, blade output

angle, and the number of blades [1]. One-dimensional Eu-

ler’s theory is usually utilized when designing the pumps.

In order to ensure one-dimensionality of the flow, infinite

number of blades is assumed with zero thickness; an ideal

assumption which is obviously impossible to reach in reali-

ty. As the number of blades reduces, the flow loses its one-

dimensionality within the area between the blades and may

no more conform from the vane; this may reduce the pump

head to a lower head than that of Euler [2, 3].

In order to increase pump head and efficiency, ra-

ther than arranging the pumps in series or parallel configu-

rations, one can use a multi-stage or multi-pressure pump

which principally works on a similar basis to that of cen-

trifugal pumps. A multi-stage pump enjoys impellers of the

same size which are designed for the same capacity. A

two-stage pump, for example, performs much like two

single-stage ones arranged in series. The first pump’s out-

put is fed into the second pump to further increase the head

at the second impeller’s output. With increasing the num-

ber of impellers increases, one may expect the ultimate

output head to rise. One of the issues arisen when using

multi-stage pumps is the cavitation phenomenon. This

phenomenon may contribute into performance losses of the

pump, corrosion of the blades and the shell, increased

noises and unwanted vibrations of the pump. Prior to use

the pump, in order to prevent potential damages to the

pump, one should forecast the behavior of the flow inside

the pump. For this purpose, before actually operating the

pump, ANSYS Fluent software is utilized to expose the

pump to three-dimensional complex flows and turbulences.

It is very expensive, time-intensive and boring to try to

analyze the flow through the experimentations and model

tests. As such, Computational Fluid Dynamics (CFD) is

usually used to have the fluid flow analyzed. In the recent

years, many industries have used this knowledge, as a

powerful numerical simulation tool, to analyze fluid flows

and centrifugal pumps.

Zhou [4] simulated the internal flow within three

types of centrifugal pumps with 4 straight blades and 12

twisted ones, in two scales. In his research, he employed a

commercial three-dimensional Navier-Stocks code, called

CFX, using a k-ε standard turbulence model. Also, he used

triangular grids to mesh the hub, mid span, and shroud

before introducing the meshed impellers into the CFX

environment. Furthermore, pressure and velocity distribu-

tion vectors corresponding to different periods were ob-

tained for all the three impellers. The research results

proved that one can enhance the pump efficiency using

curved blades rather than straight ones.

Bross et al. [5] used experimental approaches

along with numerical CFD-based methods to determine

streamlines within the shroud, hub, and on the blades of an

impeller. They investigated the effect of Reynolds number

on the outlet flow angle. They further calculated the values

of velocity and pressure across different zones around the

impellers of the pumps.

Thin Cho Thin [6] analyzed the passing flow

through a single-stage, single-suction pump. He showed

that different losses may affect the corresponding perfor-

mance curve to the pump. Hydraulic losses (due to friction,

changes in flow direction, or changes in the cross section

area of the volute casing) developed along the fluid flow

path impose significant effect on the pump flow rate, head

and power consumption. This impact may lead to increased

power consumption yet reduced the pump head. He further

illustrated that, one can rise the pump’s output head by

controlling and reducing the impeller friction losses, shock

losses, disc friction losses, volute friction losses, and recir-

culation losses. He discussed the advantageousness of the

centrifugal pumps, over other sorts of pumps, in terms of

such characteristics as higher efficiencies, continuous flow

and easier installation and maintenance protocols.

Anagnostopoulos [7] evaluated pumps perfor-

mances in terms of efficiency. Estimating total hydraulic

losses incurred across a pump’s internal and external parts

as well as its casing, he used, with the help of evaluation

software, an optimized numerical algorithm based on un-

constrained gradient to determine the pump’s maximum

mailto:[email protected]

http://dx.doi.org/10.5755/j01.mech.23.1.17676

108

efficiency. Aiming at better fitness of the meshing ele-

ments at irregular boundaries, so as to achieve results of

higher accuracy, he used advanced numerical techniques.

He showed that hydraulic losses within volute casing and

also in the pump inlet and outlet may contribute into re-

duced pump head.

Asilios [8] demonstrated that the geometry of a

blade’s leading edge may play an important role in the

design of the blade. He performed a set of tests on a 9-

blade impeller. A number of points at the blade tip were

considered as control points with the effects of changes

and displacements of these points on the pump power con-

sumption and head been demonstrated. He used averaged

Navier-Stokes equations using a k-ε standard turbulence

model with control volume approach to optimize the inter-

nal geometry of the blades of the impeller in the centrifu-

gal pump.

Kergourlay et al. [9] investigated the effects of

splitters in the design of turbo-machines. In a centrifugal

pump, too many or too few number of blades may result in

the interference of the flows inside the impeller. It is ex-

pected that, head and efficiency increase with increasing

the number of blades; however, such factors as the intensi-

fication phenomenon, blockage phenomenon due to blade

thickness, and also increased friction along the guide vane

contribute to efficiency losses. On the other hand, for a

case where the number of blades is not adequate, the fluid

may not follow one-dimensional flow principles within the

impeller with local losses (leakage) increased. Therefore,

one can see that pump performance is affected by the in-

creases in hydraulic losses. Hydraulic losses can be re-

duced using splitters. These blades (splitters) are mounted

on the central axis of the main blades to enhance absolute

velocity and total pressure.

Changing the geometry of impeller’s blades of a

centrifugal pump, Bacharoudis et al. [10] investigated the

pump performance. Their results indicated a good correla-

tion between local and global parameters. Using a finite

volume CFD program, the research provided a numerical

solution for three-dimensional incompressible Navier-

Stocks equations in an irregular grid system. Furthermore,

flow pattern and pressure distribution along the guide

vanes were calculated and flow rate-head curves were

compared and discussed.

Houlin et al. [11] investigated the contributions

from the number of blades into the specifications of a cen-

trifugal pump. Flow analyses showed that, changes in the

number of blades imposes very large effects not only on

the low-pressure zones behind the internal blade, but also

on the jet-wave structure in impellers. With increasing the

number of blades, one can expect the pump head to in-

crease as well. In this research, optimum number of blades

to achieve the maximum pump efficiency is calculated.

Furthermore, optimum number of blades to minimize the

cavitation phenomenon is calculated.

Tverdokhleb [12] proposed the use of multi-stage

pumps, rather than large-radius pumps, when a high level

of output energy was demanded. He evaluated the products

of HMS Company and used Ansys CFX to model fluid

flow, velocity distribution, and pressure distribution across

internal volumes. He demonstrated that using two-row

pumps, one can attenuate the amplitude of pressure pulsa-

tions and enhance vibration characteristics.

Using tests and LDV measurements along with

CFD techniques, Wen-Guang [13] determined fluid (water

and oil) velocity in the impeller as well as volute casing of

a centrifugal pump. He proved the slip factor to be a very

important factor when hydraulically designing impellers of

centrifugal pumps intended to work with oil. In this re-

search, the effect of fluid type (water and oil) was evaluat-

ed on the pump output head and flow rate. He showed that

the type of fluid within the impeller may directly contrib-

ute to the pump head and efficiency and shaft power.

Based on genetic algorithm optimization ap-

proach, Yang Sun-Sheng [14] undertook hydraulic design

of an impeller for a centrifugal pump. He compared six

impellers of single-arc, double-arc and triple-arc, logarith-

mic spiral and variable angle spiral to investigate the effect

of impeller trimming on the pump performance. The opti-

mization results revealed that the resulting impeller exhib-

ited better performance in terms of efficiency and required

torque to run the impeller. Also in this research, numerical

results were compared to experimental data.

Using three-dimensional Navier-Stocks equations,

Bellary et al. [15] optimized the shape of blades on the

impeller of a centrifugal pump. They incorporated CFD

into press based averaging method. In their research, the

head, efficiency, and power consumption versus flow rate

curves for three fluids: water, oil, gasoil were obtained.

They showed that the effect of flow loss across the impel-

ler’s shroud and hub may contribute increased efficiency.

Further, the effect of outlet and inlet angles on the efficien-

cy was considered and showed that, compared to inlet an-

gle; the outlet angle may have much larger contribution

into pump efficiency.

Qing Zhang et al. [16] used CFD to investigate

the effect of impeller inlet width of a centrifugal pump on

its head and efficiency levels. For different values of im-

peller inlet width, they developed pressure distribution

contours across various zones. The results indicated the

maximum pump efficiency and head to correspond to an

inlet width of 66 mm.

In the present research, using experimentations as

well as numerical simulations, the effect of the blade ge-

ometry (including blade outlet angle) of a multi-pressure

pump on the pump head and efficiency is studied. At spe-

cific values of outlet angle and blade width, hydraulic and

mechanical losses within the impeller are reduced along

with increases in the pump head and efficiency. Simulation

results exhibit very good agreement with experimental

data; Therefore, using numerical simulations with the help

of ANSYS Fluent (rather than conducting actual tests), one

can save both time and money.

2. Numerical simulation

Analyzed here is a closed single-suction, five-

blade impeller with hub and shroud coverings (see Fig. 1).

Impeller model was generated in SOLIDWORKS software

before analyzing the fluid flow using ANSYS Fluent soft-

ware. In order to compare simulation results to those of

experiments and validate the numerical solution, model

parameters and values were set to correspond to test condi-

tions. Characteristics of the impeller are reported in Ta-

ble 1.

109

Table 1

Impeller characteristics

Outlet angle,

Degree

Inlet angle,

Degree

Inlet width,

mm

Outlet width,

mm

Thickness,

mm

Outer diameter

Do, mm

Inner diameter

DI, mm

Number of

Blade

27, 30 and 33 18 42 13 4 304 125 5

Fig. 1 Pump impeller

2.1. Meshing

Cartesian grid system is used in this research.

This meshing approach is preferred over other meshing

techniques in terms of the fastness and accuracy. Devel-

opment of an adequate grid system represents the most

basic yet important step to undertake in the course of solv-

ing a problem via CFD-based approaches. Meshing plays

an important role in the convergence, required time to

achieve and the resulting accuracy of the problem solution.

Multi-block meshing method was used here. In order to

further enhance the fastness of the calculations and mesh-

ing task, the impeller was divided into five even sections

(sectors) among which one sector was analyzed (see

Fig. 2). Each sector was split into two zones. The first zone

encompasses the space around and in-between the pump

blades (moving zone), while the second zone encompasses

the remaining portion of the sector (stationary zone). In

order to provide adequate grids to boundary layer areas

surrounding the blades, moving zone, and the areas near

the blades, the grids within these areas should be set to be

as small as possible.

Fig. 2 Division of the impeller into five zones

2.2. Surface meshing

Since unstructured meshing approach have been

already shown to be well-suited, unconstrained to mesh

surfaces of the objects of complex geometries, this meshing

approach was undertaken for all surfaces in this research.

Higher speeds and lower boundary layer thicknesses ob-

served close to the blade tip implied that smaller grids

should be used across these areas, so as to consider the

effects of boundary layer. Blade surfaces were meshed with

squared elements. In order to properly model the boundary

layer and tip vortexes, the size of the square sides were set

to 0.001 Do for the blade edges, blade tip, and areas close

to the root. Other surfaces were meshed with larger square

cells of approximately 0.01 Do side lengths.

2.3. Volume meshing

In order to have volumes meshed, unstructured

meshing approach was implemented via the TGrid algo-

rithm. The generated elements by this algorithm were

quadtrial of specific sizes. This algorithm allows some

secondary elements to be generated with different sizes

than those of the primary ones. In general, in volumetric

meshing, aspect ratio parameter may not exceed 0.95, be-

cause failure to meet this criterion may lag the convergence

to the solution or even make the solution diverged. As

such, it is recommended to keep this parameter below 0.8

with its ideal value being 0.5. Here, the maximum allowa-

ble value of this parameter was set to 0.6. In addition, it is

better for each element to have a size not greater than 1.2

times the size of the adjacent element. Fig. 3 illustrates the

meshing approach for moving zone and around the blades.

110

Fig. 3 Impeller meshing

2.4. Boundary conditions

The fluid pressure at the fluid inlet of the control

volume (inlet extension) and also at the fluid outlet of the

control volume (outlet extension) was considered as the

boundary conditions. The k-ε use as the turbulence model

in the simulation. Fig. 4 presents inlet pressure and outlet

pressure boundary conditions.

Fig. 4 Pressure boundary conditions

3. Numerical results

First, pressure distribution contours were plotted

for different flow velocities (flow rates) across different

zones on the impeller. In this phase, pressure distribution

contours, pump head and efficiency values were obtained

at eight different flow rates. Using Eq. (1), pump flow rate

was determined at different flow velocities. Also, with the

help of Eq. (2), total pressure contours were used to obtain

the pump head values.

Q VA ; (1)

disch arg e suctionP P

Hg

, (2)

where H, Q, V and A denote pump head, flow rate, flow

velocity, and cross-sectional area of the flow, respectively;

ρ and g are flow density and gravitational acceleration,

respectively. Further, one can evaluate static pressure and

velocity changes for different values of flow rate and ve-

locity.

111

Continuing with the research, three impellers of

different outlet angles, namely 27°, 30°, and 33° were in-

vestigated at a rotation speed of 2000 rpm. As mentioned

before, total pressure and velocity distributions and hence

the pump head and efficiency values were then obtained at

eight different flow rates. Presented here are the corre-

sponding total pressure plots to a few flow rates only.

3.1. Total pressure contours

3.1.1. Outlet angle of 27°

Figs. 5 and 6 demonstrate total pressure contours

corresponding to inlet velocities of 0.68 m/s (equivalent to

a flow rate of 500 l/min) and 2.71 m/s (equivalent to a flow

rate of 2000 l/min), respectively.




a flow rate of 1500 l/min) and 2.71 m/s (equivalent to a

flow rate of 2000 l/min), respectively.




a flow rate of 500 l/min) and 2.03 m/s (equivalent to a flow

rate of 1500 l/min), respectively.

Fig. 5 Total pressure contours corresponding to the inlet velocity of 0.68 m/s; outlet angle: 27°


112




113


3.2. Head-flow rate performance curves

Using the pressure contours and Eq. (2), corre-

sponding pump head values to different flow rates were

obtained. Where, pressure values are obtained from Fig. 7.

Table 2 reports corresponding pump head values to eight

different flow rates at each of the three outlet angles (27°,

30°, and 33°).

Table 2

Numerical values of pump head for eight different flow rates; outlet angles: 27°, 30°, and 33°

Q, lit/min H, m

27 CFD 30 CFD 33 CFD

500 42.8 58.5 54.8

1000 41.8 57.8 53.8

1500 38.3 55.3 49.1

2000 32.4 51.4 40.2

2500 27.1 45.8 32.1

3000 17.6 36.7 22.6

3230 11.2 27.4 19.9

3500 6.9 19.8 12.7

Plotted in Fig. 11 are the head-flow rate perfor-

mance curves at each of the three outlet angles (27°, 30°

and 33°). CFD analysis show that with the adjustment of

the impeller geometry of the pump, the head and efficiency

at 30 degree outlet angle in comparison with other cases

increase.

Fig. 11 Head-flow rate performance curves at each of the three outlet angles (27°, 30°, and 33°)

114

3.3. Efficiency-flow rate performance curves

Pump efficiency can be calculated from the

Eq. (3):

gHQ

P

, (3)

where, P denotes the pump input power (here 30 kW) and

η represents the pump efficiency. Reported in Table 3 are

the values of efficiency in terms of flow rate for the three

outlet angles.

Plotted in Fig. 12 is the efficiency-flow rate per-

formance curves at each of the three outlet angles (27°,

30°, and 33°). CFD analysis show that the efficiency of

pump in 2500 (lit/min) is maximum for all type of outlet

angle blades and for 30 degree outlet angle is better than

others.

Table 3

Numerical values of pump efficiency for eight different flow rates; outlet angles: 27°, 30°, and 33°

Q, lit/min η, %


500 11.86 16.25 15.02

1000 23.21 32.11 29.88

1500 31.91 40.08 40.83

2000 37.11 56.88 44.66

2500 37.51 63.61 45.61

3000 29.3 61.16 37.66

3230 20.02 49.16 35.71

3500 13.41 38.51 24.69

Fig. 12 Efficiency-flow rate performance curves at each of the three outlet angles (27°, 30°, and 33°)

4. Test

A multi-stage pump of 2000 rpm and 30 kW input

power was used for the tests. The pump impeller was man-

ufactured according to the specifications in Table 1.

Figs. 13 and 14 demonstrate the used multi-stage pump

and its impellers, respectively. For achieve the results of

different outlet angle, initially according Fig. 13 disassem-

ble a multi stage pump device and bring out the low pres-

sure impeller from original shaft. After assembly the im-

peller with 27 degree outlet angle on the pump and install

them on the chassis of midlum truck device. Then with set

up the three parts of propeller shaft (one part fix and two

part moveable) transfer the rotation of 2000 rapid per mi-

nute from motor to pump with power take off using the

rapid meter for adjust the revolution that mounted on con-

trol panel. We start testing the pump in different flow rate

with opening the inlet valve that adjoined to water tank. At

first install hose on the outlet then open the outlet valve

and adjust the flow meter on 500 liter per minute. After

start measure the head of the pump and note the results.

Respectively repeat this for another flow rate (1000, 1500,

2000, 2500, 3000, 3500 lit/min). In the next step disclose

pump from chassis and after disassemble, assembly the

impeller with 30 degree outlet angle. Similar to first stage

note the results the head of the pump in different flow rate.

we will do all cycle for 33 degree outlet angle and note the

outlet head and in next level compare these with CFD re-

sults.

Fig. 15 provides a schematic of test setup and

testing machine. Pump suction the water from the tank and

shooting the water then we can meter the head of pump in

various outlet angle blade.

115

Fig. 13 The used multi-stage pump for the experimentations

Fig. 14 Pump impellers with different outlet angles: 27°, 30°, and 33°

Fig. 15 The schematic of test setup

5. Experimental results

5.1. Head–flow rate performance curves

Table 4 reports corresponding experimental pump

heads to eight different flow rates at each of the three out-

let angles (27°, 30°, and 33°).

Plotted in Fig. 16 are the experimentally derived

head-flow rate performance curves at each of the three

outlet angles (27°, 30°, and 33°). The test show that with

the adjustment of the impeller geometry of the pump, the

head and efficiency at 30 degree outlet angle in compari-

son with other cases increases.

116

Table 4

Experimental values of pump head for eight different flow rates; outlet angles: 27°, 30°, and 33°

Q, lit/min H, m

27 Exp 30 Exp 33 Exp

500 40.3 55.1 53.2

1000 38.8 54.5 52.5

1500 36.8 50.2 48.5

2000 31.9 45.5 42.8

2500 26.2 41.5 35.1

3000 20.1 29.8 25.8

3230 10.3 22.2 17.4

3500 5.3 15.1 12.2

Fig. 16 Experimentally derived head-flow rate performance curves at each of the three outlet angles

5.2. Efficiency-flow rate performance curves

Table 5 reports corresponding experimental pump

efficiencies to eight different flow rates at each of the three

outlet angles (27°, 30°, and 33°).

Plotted in Fig. 17 are the experimentally derived

efficiency-flow rate performance curves at each of the

three outlet angles (27°, 30°, and 33°).

Fig. 17 Experimentally derived efficiency-flow rate performance curves at each of the three outlet angles

117

6. Simulation results versus experimental data and

CFD model validation

Figs. 18 and 19 present the comparisons between

the numerical and experimental results in terms of head-

flow rate and efficiency-flow rate performance curves,

respectively. The CFD analysis and experimental results

show that with the adjustment of the impeller geometry of

the pump, the head and efficiency at 30 degree outlet angle

in comparison with other cases increases and improves the

centrifugal pump performance. Numerical and experi-

mental results show that the efficiency of the pump in

2500 lit/min is maximum relative to the other cases.

Fig. 18 Head-flow rate performance curve, outlet angles: 27°, 30°, and 33°

Fig. 19 Efficiency-flow rate performance curve, outlet angles: 27°, 30°, and 33°

Table 5

Experimental values of pump efficiency for eight different flow rates; outlet angles: 27°, 30°, and 33°

Q, lit/min η, %


500 11.11 15.27 14.72

1000 21.55 30.27 29.16

1500 30.66 41.66 40.41

2000 35.55 50.55 47.77

2500 36.11 52.08 48.61

3000 33.33 50.12 42.87

3230 17.94 39.47 31.22

3500 9.77 29.16 23.72

118

7. Conclusion

The effect of outlet angle geometry of multi-

pressure pump impellers on the pump head and efficiency

were investigated both numerically (via CFD) and experi-

mentally. A closed single-suction, five-blade impeller was

analyzed at three outlet angles: 27°, 30°, and 33°. First, the

impeller model was developed in SOLIDWORKS soft-

ware; then it was used to analyze the fluid flow using AN-

SYS Fluent software. Unstructured meshing approach was

followed for all surfaces and volumes. Blade surfaces were

meshed with square elements while quadtrial elements

were used to mesh the volumes. In the experimental part of

the research, tested with three impellers of different outlet

angles was a multi-stage pump of rotation speed and input

power of 2000 rpm and 30 kW, respectively. Numerical

simulation results were in good agreement with the test

data, particularly at outlet angles of 27° and 33°. There-

fore, one can save money and time by using numerical

simulation tools provided by ANSYS Fluent, rather than

undertaking experimental approaches.

Both numerical and experimental results indicated

that, at specific values of outlet angle and blade width,

hydraulic and mechanical losses within the impeller re-

duce, increasing the pump head and efficiency. Maximum

pump head and efficiency were observed to occur at an

outlet angle of 30°. This is because of the reduced losses

due to water recirculation within the volute casing and

outlet impellers. At this angle, improved energy transfer

into the fluid (energy consumption) and better conductance

of the flow by the main blades in the impeller outlet may

result in increased pump head and efficiency.

Conflict of interests

The authors declare that there is no conflict of in-

terests regarding the publication of this paper.

Refrences

1. Gülich, J.F. 2010. Centrifugal Pumps, Springer-Verlag

Berlin Heidelberg, Second edition, 966 p.

http://dx.doi.org/10.1007/978-3-642-12824-0.

2. Nourbakhsh, A. 2005. Turbo Machines, Tehran uni-

versity publications, Second edition.

3. Stepanoff, A.J. 1993. Centrifugal and Axial Flow

Pumps: Theory, Design and Application, Krieger Pub-

lishing Company, Second edition, 470 p.

4. Zhou, W.; Zhao, Z.; Lee, T.S.; Winoto, S.H. 2003.

Investigation of flow through centrifugal pump impel-

lers using computational fluid dynamics, International

Journal of Rotating Machinery 9(1): 49-61.

http://dx.doi.org/10.1155/S1023621X0300006X.

5. Hamkins, C.P.; Bross, S. 2002. Use of surface flow

visualization methods in centrifugal pump design,

Journal of Fluids Engineering 124(2): 314-318.

http://dx.doi.org/10.1115/1.1470477.

6. Thin, K.C.; Khaing, M.M.; Aye, K.M. 2008. Design

and performance analysis of centrifugal pump, World

Academy of Science Engineering and Technology

46: 422-429.

7. Anagnostopoulos, S. 2009. A Fast Numerical Method

for Flow Analysis and Blade Design in Centrifugal

Pump Impellers, Computers & Fluids 38(2): 284-289.

http://dx.doi.org/10.1016/j.compfluid.2008.02.010.

8. Grapsas, V.A.; Anagnostopoulos, J.S.; Papantonis, D.E. 2005. Hydrodynamic design of radial flow pump

impeller by surface parameterization, 1st International

Conference on Experiments/Process System Model-

ing/Simulation/Optimization, Athens, 6-9 July.

9. Kergourlay, G.; Younsi, M.; Bakir, F.; Rey, R. 2007.

Influence of splitter blades on the flow field of a cen-

trifugal pump: test-analysis comparison, International

Journal of Rotating Machinery 7: 1-13.

http://dx.doi.org/10.1155/2007/85024.

10. Bacharoudis, E.C.; Filoios, A.E.; Mentzos, M.D.;

Margaris, D.P. 2008. Parametric study of a centrifugal

pump impeller by varying the outlet blade angle, The

Open Mechanical Engineering Journal 2(1): 75-83.

http://dx.doi.org/10.2174/1874155X00802010075.

11. Houlin, L.; Yong, W.; Shouqi, Y.; Minggao, T.; Kai,

W. 2010. Effects of blade number on characteristics of

centrifugal pumps, Chinese Journal of Mechanical En-

gineering 23(6).

12. Tverdokhleb, I.; Knyazeva, E.; Birukov, A.;

Lugovaya, S. 2010. About designing the flow part of a

multi-stage pump with a minimum radial dimensions,

Procedia Engineering 39: 84-90.

http://dx.doi.org/10.1016/j.proeng.2012.07.011.

13. Li, W.G. 2013. Effect of flow rate and viscosity on slip

factor of centrifugal pump handling viscous oils, Inter-

national Journal of Rotating Machinery Article ID

317473.

http://dx.doi.org/10.1155/2013/317473.

14. Zhou, X.; Zhang, Y.; Ji, Z.; Hou, H. 2014. The opti-

mal hydraulic design of centrifugal impeller using ge-

netic algorithm with BVF, International Journal of Ro-

tating Machinery Article ID 845302.

http://dx.doi.org/10.1155/2014/845302.

15. Bellary, S.A.L.; Samad, A. 2014. Improvement of

efficiency by design optimization of a centrifugal pump

impeller, ASME.

http://dx.doi.org/10.1115/gt2014-25217.

16. Zhang, Q.; Zhou, H.; Gao, Q.; Cui, Z. 2014. Analy-

sis of effects of impeller inlet width on the performance

of centrifugal pump, Journal of Chemical and Pharma-

ceutical Research 6(5): 2078-2081.

N. Mohammadi, M.Fakharzadeh

ANALYSIS OF EFFECT OF IMPELLER GEOMETRY

INCLUDING BLADE OUTLET ANGLE ON THE

PERFORMANCE OF MULTI-PRESSURE PUMPS:

SIMULATION AND EXPERIMENT

S u m m a r y

In this paper, the effect of outlet angle geometry

of the blades of the impeller of a multi-pressure pump on

the pump’s head and efficiency using two approaches,

namely numerical approach (Computational Fluid Dynam-

ics) and experimentation are investigated. Herein, a closed

single-suction, five-blade impeller with three different

outlet angles (27°, 30°, and 33°) is analyzed. The impeller

model was first built in SOLIDWORKS software before

analyzing the fluid flow with the help of ANSYS Fluent

software. All surfaces were meshed through irregular grid-

119

ding. Considering high flow velocity along with low thick-

ness of boundary layer near the blade tips, smaller mesh

sizes were used across these areas. Blade surfaces were

meshed using square elements. Furthermore, in order to

model the volumes, irregular gridding and TGrid algorithm

were undertaken. Generated with this algorithm are

quadtrial elements of known size. In volumetric meshing

stage, aspect ratio was set to a maximum of 0.6. In the

experimentation phase of this research, a multi-stage pump

was tested at 2000 rpm and 30 kW of input power with

three impellers of different angles. At specific values of

outlet angle and impeller outlet width, some reductions in

hydraulic and mechanical losses within the impeller along

with some rises in the pump head and efficiency were ob-

served. Simulation results were found to be in good

agreement with the experimental data; so as one can save

time and money by using numerical simulation approaches

in ANSYS Fluent rather than undertaking experimenta-

tions. Numerical and experimental analyses revealed that

the maximum pump head and efficiency have been wit-

nessed at the outlet angle of 30°. This is because of the

reduced losses by water recirculation within the volute and

outlet impellers.

Keywords: Impeller Geometry, centrifugal pump, ANSYS

Fluent.

Received October 05, 2016

Accepted February 06, 2017

analysis of effect of impeller geometry including blade

Documents